Demo plan

• configure

• local llm reply

• cloud llm reply

• extra:

  • ?

• leuke bug

  • bug die pas op te lossen

Demo setup

Setup cluster

kind create cluster

kubectl apply -f ./k8s-bugs/

Ollama on Ubuntu

Recommended: enable usage of gpu inside ollama container

# Setup Ollama
ollama pull llama3.1

docker run -it --rm \
  --device /dev/dri \
  -e DISPLAY=$DISPLAY \
  -v /tmp/.X11-unix:/tmp/.X11-unix \
  -v ollama:/root/.ollama \
  -e OLLAMA_DEBUG=1 \
  -p 11434:11434 \
  --name ollama \
  ollama/ollama

docker exec -it ollama ollama run llama3.1

# Configure k8sgpt with ollama backend
k8sgpt auth add --backend ollama --baseurl http://localhost:11434 --model llama3.1

# Make it the default backend
k8sgpt auth default --provider ollama

k8sgpt analyze --explain

Troubleshooting

cat ~/.config/k8sgpt/k8sgpt.yaml

Use your gpu

Mesa Intel Graphics

On Ubuntu

# Install Mesa / gpu tools
sudo apt install mesa-utils intel-opencl-icd
sudo apt install intel-gpu-tools intel-media-va-driver-non-free

# Allow Access to X Server
xhost +local:

# Verify
docker run -it --rm \
    --device /dev/dri \
    -e DISPLAY=$DISPLAY \
    -v /tmp/.X11-unix:/tmp/.X11-unix \
    ubuntu bash

ls /dev/dri
apt update
apt install mesa-utils
glxinfo | grep "OpenGL renderer"